The starting line of last weekend’s The North Face 50 Mile Championship was a who’s-who of the ultra running scene, primarily domestic but with some international flavor, too. Some of the faces I knew from Boulder or past industry events like Galen Burrell, Sage Canady, Dave Mackey, and Jorge Maravilla. But others I knew only through iRunFar.com and Strava, like Dylan Bowman, Rob Krar, Timmy Olson, and Alex Varner.
Every year, the TNF50 field seems to claim the title of being the deepest and most talented field ever. The elites are drawn by the opportunity to win significant prize money, to receive great exposure, and/or to compete against some of the other best ultra runners in the world. Plus, the event has a near monopoly of the calendar, unlike the summer months when runners are scattered among similarly scheduled races.
In this post I’d like to examine how this year’s field compared to those in the past.
Most talented field ever?
I’ll first take on the issue of talent since I’m going to take a pass on it. This is a conversation that could happen elsewhere among more insightful commentators. But I don’t feel comfortable comparing the winners this year to those in past years, e.g. “Would the Sage of 2014 have beaten the Krar of 2013 or the Heras of 2012 or 2010”?
If the course and course conditions were identical each year, I might be more willing, but that’s rarely been the case. Without that constant, comparisons are very subjective.
Deepest field ever?
I don’t mean to reinforce my reputation as “the chart guy,” but here we go again…
By examining past results, the depth of a field can be measured. By “depth,” I mean the difference in finishing times between the top finishers and later finishers. In a deep field, little time will separate first place from, say, 30th place. In this type of race, small mistakes — like a shoelace that becomes untied, temporary GI discomfort, or a disorganized aid stop — can cost precious places.
1. I retrieved finishing times for the Top 50 finishers for the past five TNF50 races:
2. I converted the times into decimal hours. For example, Miguel Heras’ 2010 winning time of 6:47:03 converts to 6.78.
3. I calculated the finishing times of the Top 50 as a proportion of the winner’s time. For example, if Heras had finished in exactly 6.0 hours and the 50th place finisher crossed in 9.0 hours, the 50th finisher’s proportional time would be 1.50. (The winner’s proportional time is always 1.0.)
If I had simply used the raw time difference between the winner and the other finishers, the races with relatively fast conditions — like the shortened course in 2012 — would appear deeper than races with relatively slow conditions, simply because there was less time for the runners to “spread out.”
By using proportional time, I realize that I’m assuming that all winning efforts were equal. This may not have been the case but I’m not going there — see the “Most talented field ever?” section above.
Here is my chart of the Top 50 finishers for the TNF50 in years 2010 through 2014:
The data is very clear: at least if measured by the Top 50, the 2014 field was the deepest ever. The rest of the order: 2011, 2013, 2012, and finally 2010.
I zoomed in more closely on the Top 25, since the Top 50 chart does not clearly show what is happening there:
If we only look at the Top 25, the 2011 field was actually the deepest, just ahead of this year. But really, 2011-2014 are all very close; the only year that stands out is 2010, when the field dropped off quickly starting with the 12th finisher.
Finally, I looked at just the Top 10:
I’d be reluctant to define the depth of a 350-person race based on Top 10 data. Instead, it really just shows how competitive — or not competitive — the fight at the front was. In 2010 and 2014, there was a great fight for the top 5. In 2011, Mike Wolfe and Dakota Jones were hammering each other for the win. And in 2012 and 2013, the race was dominated by Miguel Heras and Rob Krar.
Nice work! The talent and training levels are really so close, the charts are almost necessary. For instance, Kelly Slater staying near the top of surfing for 20 years, that would be impossible in ultras, so great is the simple feat.
I’m also interested in longevity for professional ultra runners. If optimal health was the goal, how many long races are sustainable in a year? In 3 years? In 5? Is fitness (Vo2 / strength / flexibilty) actually improving throughout a racing career? Or do the demands of the distance x terrain actually have a detrimental effect over time…
Can you answer that with a single chart please? By EOD if possible.
To expand a bit on your analysis, I ran a linear regression in excel to come up with an estimate of the average increase in proportional time per place for the first 25 finishers over the last 5 races. According to my numbers, 2014 was the deepest year through the 1st 25 places, and 2012 was the least deep (shallowest?). I should point out that I didn’t test for statistical significance, but judging from the standard deviations, it doesn’t look to me like any of these differences are actually significant. Anyway, here are my results:
2014–(mean: .0090, SD: .0018)
2013–(mean: .0099, SD: .0040)
2012–(mean: .0127, SD: .0071)
2011–(mean: .0103, SD: .0039)
2010–(mean: .0109, SD: .0010)
I forced the line to go through the point (1,1), meaning the first place finisher always has a proportional time of 1. Because of this, the slope of the regression line is the only thing we need to estimate.
See, this is why I don’t understand my new reputation as “the chart guy.” I’m just a statistical hack; you obviously know it on another level.
I’m surprised by your finding that 2012 was the least deep field. My charts show a different story — in 2010, the field drops off quickest. What am I missing?
If you compare 2010 to 2012, and look at where the biggest drop-off in times happened, you’ll notice that in 2012 the largest drop-off happens in the top 10. In 2010, the large drop-off doesn’t happen until after 10th place finisher. This is the reason for the discrepancy: the method I used will naturally weight the top finishers more heavily than the lower finishers.
Now whether you agree that that is an accurate estimate of the depth of the field could be up for debate, but that’s the mathematical reason behind that result.